Use fixed thread pool

There was no reason for using a cached thread pool. The number of tasks
(partion solvers, partCount) is fixed. They all need to be executed in
parallel to prevent partition starving, therefore the number of threads
needed is exactly same as parrtCount. No additional threads are needed.

optaplanner-core/src/main/java/org/optaplanner/core/impl/partitionedsearch/DefaultPartitionedSearchPhase.java

@@ -20,8 +20,8 @@
 import java.util.List;
 import java.util.ListIterator;
 import java.util.concurrent.ExecutorService;
+import java.util.concurrent.Executors;
 import java.util.concurrent.Semaphore;
-import java.util.concurrent.SynchronousQueue;
 import java.util.concurrent.ThreadFactory;
 import java.util.concurrent.ThreadPoolExecutor;
 import java.util.concurrent.TimeUnit;
@@ -174,11 +174,8 @@ public void solve(DefaultSolverScope<Solution_> solverScope) {
     }
 
     private ExecutorService createThreadPoolExecutor(int partCount) {
-        // Based on Executors.newCachedThreadPool(...)
-        ThreadPoolExecutor threadPoolExecutor = new ThreadPoolExecutor(0, Integer.MAX_VALUE,
-                60L, TimeUnit.SECONDS,
-                new SynchronousQueue<>(),
-                threadFactory);
+        ThreadPoolExecutor threadPoolExecutor
+                = (ThreadPoolExecutor) Executors.newFixedThreadPool(partCount, threadFactory);
 
         if (threadPoolExecutor.getMaximumPoolSize() < partCount) {
             throw new IllegalStateException(